There exists a need for improved substrates for component parts of head-disc assemblies (also known as "Winchester drives"), particularly substrates for hard discs and read/write heads.
Currently, the most widely used substrate material for hard discs and read/write heads is aluminum or aluminum alloys. To produce a hard-disc from aluminum or an aluminum alloy, an aluminum or aluminum alloy substrate is blanked, thermal flattened, sized and chamfered and diamond-turned to a surface roughness of about 250 .ANG. RMS. The blank is then chemically pretreated to remove aluminum oxide prior to coating with an undercoat layer. The undercoat layer is typically electroless nickel/phosphorous containing sufficient phosphorous so as to be non-magnetic. The nickel/phosphorous layer is polished and then textured. To the undercoat is then applied a magnetic coating, typically a cobalt/phosphorous alloy. Finally, a protective overcoat is applied, e.g., sputtered carbon. Additional layers and alternative procedures may be used, depending upon the manufacturer.
There is a continuing desire to achieve higher memory capacity on discs, including higher storage (bit) density and track density. Aluminum and aluminum alloy substrates have inherent limitations with respect to achieving higher storage and track density; hence, there is a desire for substitute materials. Among aluminum's and aluminum alloys' limitations with respect to their use as a substrate where high storage density is required are low elastic modulus, high coefficient of thermal expansion, and low Knoop hardness. Furthermore, aluminum exhibits poor chemical resistance, oxidation resistance, thermal stability and polishability. The poor polishability (limited to about 100 .ANG. RMS) necessitates the nickel/phosphorous undercoat.
Several materials have been considered as alternatives to aluminum, as for example, as discussed in U.S. Pat. No. 4,808,463, the teachings of which are incorporated herein by reference.
Glasses have certain advantages relative to aluminum and aluminum alloys, e.g., very low coefficient of thermal expansion, but also have limitations as substrate for magnetic recording components; in particular, being electrically non-conductive, and having very low thermal conductivity (watts per meter per .degree.Kelvin (W/mK)).
Silicon carbide has a number of inherent properties which suggest its use as substrates for magnetic recording components, in particular, high specific stiffness, strength, hardness, thermal conductivity, low thermal expansion and chemical and oxidation resistance and is electrically conductive.
Silicon carbide produced by sintering (e.g., Japanese patent document 88-128885/19, Sep. 12, 1986 Hitachi KK) and reaction bonding (e.g., U.S. Pat. No. 4,598,017, the teachings of which are incorporated herein by reference) have been tested for use in magnetic recording media. To produce sintered silicon carbide, powdered silicon carbide is admixed with sintering aids and compacted using heat and pressure. The need for sintering aids results in the sintered silicon carbide having voids (sintered silicon carbide generally has a density of no greater than about 90% of theoretical density), impurities (residues of the sintering aids), and has relatively loosely bound crystals. Accordingly, sintered silicon carbide must be coated if it is to be used as a substrate for recording media. In reaction bonded silicon carbide, silicon fills the voids; thus reaction bonded silicon carbide is heterogeneous, the silicon detracting from desirable properties of silicon carbide. The best polishability of reaction-bonded SiC is about 30-50 .ANG. RMS; sintered SiC is even less polishable.
It has been earlier proposed by others to evaluate silicon carbide which is deposited by chemical vapor deposition (CVD) as a candidate material for substrates of magnetic recording components. CVD-produced silicon carbide can closely approach 100% of theoretical density, has a tightly bound granular structure, and good polishability. Because CVD-produced SiC is highly polishable, it can be directly coated with a magnetic coating media, unlike aluminum or aluminum alloys which require application of a nickel/phosphorous undercoat prior to application of the magnetic coat.
Silicon carbide is generally deposited by CVD from a gaseous mixture of methyltrichlorosilane (MTS), H.sub.2, and generally an inert or non-reactive gas such as argon, helium or nitrogen, argon being preferred. Free-standing SiC is typically pyrolytically deposited on a mandrel, such as a graphite mandrel, from which it is removable. The MTS is the preferred source of both the Si and C and provides these in stoichiometric (1:1) ratios. The H.sub.2 scavenges Cl, producing HCl. The inert or non-reactive gas acts as a carrier gas for MTS (which is liquid at ambient temperatures); can be varied to adjust velocity of gas flow through the furnace as is necessary to sweep reaction product, such as HCl, from the deposited SiC; and acts as a diluent, preventing gas-phase reactions which might introduce impurities into the SiC.
CVD production of free-standing SiC material by providing a furnace having a deposition chamber and a mandrel therein and pyrolytically depositing SiC on the mandrel are described, for example, in U.S. Pat. Nos. 4,900,374; 4,997,678; and 5,071,596, the teachings of these patents being incorporated herein by reference.
The present invention is directed to CVD-deposited silicon carbide particularly suitable for magnetic recording media, particularly components of a head-disc assembly (HDA). The CVD produced in accordance with the invention has a combination of excellent thermal conductivity and high polishability not heretofore achieved in free-standing silicon carbide.
The fabrication of HDAs, including hard discs and read/write heads are described, for example, in R. W. Wood, "Magnetic Recording Systems", Proc. of the IEEE, 74(11), 1557-1569 (1986); C. Warren, "Rigid-disk Drives: Capacity, Performance Mount as Size Shrinks", Electronic Design, Apr. 28, 1983, pp. 139-150; Ivan Flores, "Chapter 5: External Storage" in The Professional Microcomputer Handbook (Van Nostrand Reinhold Co., New York, N.Y., 1986) pp. 111-151; U.S. Pat. No. 4,647,494; and articles in IBM Disc Storage Technology February 1980, including "Film Head Development" by D. A. Thompson et al. (pp. 3-5) and "IBM 3370 Film Head Design and Fabrication" by R. E. Jones, Jr. (pp. 6-9), the teachings of each of these being incorporated herein by reference.
High polishability is a very important attribute of a substrate for magnetic recording components of HDAs or the like. The a real density storage of a magnetic disc is determined by the size of the individual magnetic domains which can be achieved. The magnetic domain (or cell) size is directly related to the head fly height, i.e., the distance or gap by which the read/write head "flys" above the hard disc. The lower the fly height, the smaller the magnetic domains that can be achieved. Therefore, lower fly height translates into higher a real density. Fly height is controlled by a number of factors, surface smoothness being one of them. A limiting factor of the gap between the head and the disc (fly height) appears to be surface roughness. Surface roughness produces turbulence in the "air" gap between the disc and head, causing the head to crash into the disc if the gap is too narrow. Furthermore, lower fly heights translate to higher track density on the hard disc, increasing the overall memory storage capacity of the disc.